360 Secure SkillHub
360 安全技能中心

seo-backlink-gap

该 Skill 通过 SE Ranking MCP 拉取目标站与 3–5 个竞争对手的推荐域名,按 Domain Trust、dofollow 和交叉数量过滤并剔除已链接域名,再用 WebFetch 做主题相关性评分,输出含权威值、锚文本样本、链接类型与外联角度的潜客清单和报告。

未知作者
热度 0.00
最近更新 2026-10-18 21:34:49
分类:营销依赖:MCP 服务 / 联网浏览搜索 / 读写文件应用场景:市场分析 / 报告撰写 / 研究

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add seoskills.sh/seo-backlink-gap
Metadata
name
seo-backlink-gap
description
Find referring domains that link to multiple competitors but not to your site, then enrich with authority, anchor samples, and outreach angle per row. Produces a prospect list an outreach team can start emailing tomorrow. Use when the user asks for backlink gap analysis, link building opportunities, competitor backlink intersection, link prospecting, or wants referring domains they are missing.

Example output: examples/seo-backlink-gap-linear-app-20260514/REPORT.md

Backlink Gap

Produce an actionable link-prospecting list: domains linking to your top competitors but not to you, filtered by authority and relevance, enriched with anchor samples and a suggested outreach angle.

Prerequisites

  • SE Ranking MCP server connected.
  • Claude's WebFetch tool available (used for prospect scoring and topical relevance checks).
  • User provides: (a) target domain, (b) 3 to 5 competitor domains, and optionally (c) minimum Domain Trust (default: 25), (d) dofollow-only filter (default: true), (e) minimum intersection count (default: linked by at least 2 of the N competitors).

Process

  1. Baseline target backlinks DATA_getBacklinksRefDomains, DATA_getBacklinksSummary

    • Pull the target domain's existing referring domains to build an exclusion set.
    • Save count and summary metrics for the report.
  2. Per-competitor referring domains DATA_getBacklinksRefDomains

    • For each competitor, pull referring domains with dofollow_only=true and min_domain_trust={threshold}.
    • Collect as a dict keyed by referring domain with the list of competitors linking to it.
  3. Intersection

    • Retain referring domains that link to at least the minimum number of competitors.
    • Exclude any domain already in the target's backlink set.
  4. Enrichment DATA_getBacklinksAuthority, DATA_getBacklinksAnchors

    • For each candidate, pull: Domain Trust, Page Trust of the linking page, linking-to-competitor anchor samples (top 3 anchors), and country if available.
    • Classify by link type: editorial, resource list, directory, forum/UGC.
  5. Relevance scoring

    • Score each candidate on: (a) topical overlap (use domain homepage title/meta via WebFetch), (b) DT, (c) intersection count, (d) link-type preference.
    • Suggest an outreach angle per row: local fit, topical fit, competitive parity, resource-list inclusion, broken link, etc.
  6. Produce the prospect list (see Output Format).

Output format

Create a folder seo-backlink-gap-{target-slug}-{YYYYMMDD}/ with:

seo-backlink-gap-{target-slug}-{YYYYMMDD}/
├── 01-target-baseline.md
├── 02-competitor-{domain}-refs.md   # one per competitor
├── 03-intersection-raw.md
├── prospects.csv                     # machine-readable output
└── REPORT.md

REPORT.md follows this shape:

# Backlink Gap: {target} vs {competitors}

## Filters applied
- Dofollow only: yes
- Min Domain Trust: {n}
- Min intersection: links to at least {n} of {total} competitors

## Top 25 prospects

| # | Referring domain | DT | Links to | Sample anchor | Link type | Angle | Score |
|---|---|---|---|---|---|---|---|
| 1 | example.com | 68 | {3 of 5} | "best yoga studios" | Resource list | Topical fit | 92 |
| 2 | ... | ... | ... | ... | ... | ... | ... |

## Outreach brief
- Segment these prospects into 3 batches by link type and pitch each batch with a tailored template.
- For editorial links: pitch a unique-angle comparison or original research.
- For resource lists: suggest inclusion with a specific anchor.
- For directories: apply directly.
- For forums/UGC: engage before pitching; these are not cold-outreachable.

## Counts
- Total unique referring domains across competitors: {n}
- Passed filters: {n}
- Already linking to {target}: {n} (excluded)
- Final prospects: {n}

## Files
- prospects.csv: import to Hunter.io, Pitchbox, or your outreach tool
- Per-competitor raw data: 02-competitor-*.md

prospects.csv columns: rank,referring_domain,domain_trust,links_to_count,links_to_competitors,sample_anchor,link_type,outreach_angle,score

Tips

  • Data API rate limit: 10 requests per second. With 5 competitors, expect 5-10 seconds of sequential API time plus enrichment.
  • Do not include subdomains of already-linked domains in the prospect list. If news.example.com links to the target but blog.example.com links only to competitors, treat as already-linked unless the user asks otherwise.
  • The "Angle" column is the most important output for outreach. Keep it specific: "Their 'best X tools' list from 2024 is out of date and doesn't include you" beats "topical fit".
  • Exclude obvious spam/low-quality domains even if they pass DT thresholds. Use the WebFetch step to sanity-check any prospect scoring above 80.
  • If the target already outranks all competitors for branded terms, prioritise non-branded editorial placements over parity directories.